eval-fsr-a1-stack-ruby-swe-r469-rf0710-traces
收藏资源简介:
该数据集记录了AI代理或模型在执行任务过程中产生的交互轨迹与元数据。数据以结构化形式存储,包含对话记录、执行代理、模型信息、任务标识、运行结果和验证输出等字段。具体特征包括:对话内容(conversations,其中每个条目包含角色和文本内容)、代理类型(agent)、模型名称与提供商(model, model_provider)、日期(date)、任务类型(task)、运行批次标识(episode, run_id, trial_name)、执行结果(result)、验证器输出(verifier_output)以及数据来源(trace_source)。数据集规模为4964个样本,适用于AI代理行为分析、任务执行轨迹研究、模型评估与验证等场景。
This dataset captures the interaction trajectories and metadata generated by AI Agents or models during task execution. The data is stored in a structured manner, encompassing fields including conversation records, executing agents, model information, task identifiers, operational results, and verification outputs. Its specific features are as follows: conversations (each entry contains a role and its corresponding text content), agent type, model name and provider (model, model_provider), date, task type, operational batch identifiers (episode, run_id, trial_name), execution results (result), verifier outputs (verifier_output), and data source (trace_source). The dataset consists of 4964 samples in total, and is suitable for scenarios such as AI Agent behavior analysis, task execution trajectory research, model evaluation and verification.
- 数据集名称:laion/eval-fsr-a1-stack-ruby-swe-r469-rf0710-traces
- 数据集来源:Hugging Face Datasets(laion 组织)
- 数据集大小:
- 数据集总大小:997,760,484 字节(约 952 MB)
- 下载大小:621,074,814 字节(约 592 MB)
- 数据分割:仅包含训练集(train),共 4,964 个样本
- 数据特征:
conversations(对话列表):包含content(字符串类型)和role(字符串类型)agent(字符串):智能体标识model(字符串):模型名称model_provider(字符串):模型提供方date(字符串):日期task(字符串):任务描述episode(字符串):轮次run_id(字符串):运行 IDtrial_name(字符串):试验名称result(字符串):结果verifier_output(字符串):验证器输出trace_source(字符串):溯源来源
- 数据格式:数据文件路径为
data/train-*,包含多个文件 - 用途:该数据集主要用于评估或验证某些人工智能模型在特定任务上的表现,包含对话记录、模型输出、验证结果等结构化信息。




